ISI - International Statistical Institute Newsletter Volume 26, No. 1 (76) 2002
New faces at the ISI Permanent Office
The ISI office staff has been augmented by several new personnel
Sieriel Hoesenie: Sieriel will be processing ISI publications orders, and assisting with the general ISI accounting.
Peter von Vaupel Klein: Peter will be taking responsibility for the ISI's complex bookkeeping and accounting tasks.
Sunny Nabialczyk: Sunny will take charge of the ISI graphic design and typesetting as well as a range of other administrative tasks such as processing the outgoing post and general filing. Sunny will also be familiarising herself with the design and maintenance of the ISI WebPages.
Mention should also be made of Margaret de Ruiter-Molloy. Although hardly a new face at the ISI (having joined the ISI Permanent Office staff in 1999), Margaret has increased her working hours and responsibilities and will now be in charge of the ISI membership database.
|
Deceased Members |
|||
|
The ISI regrets to announce the death of our colleagues: |
|||
| Born | Elected | Deceased | |
| Prof. Ralph A. Bradley | 1923 | 1969 | 30 Oct. 2001 |
| Dr. Jeanne E. Griffith | 1950 | 1992 | 3 Aug. 2001 |
| Prof. Dr. Shanti S. Gupta | 1925 | 1973 | 11 Jan. 2002 |
| Prof. Henry O. Lancaster | 1913 | 1961 | 2 Dec. 2001 |
| Prof. Harold Ruben | 1923 | 1974 | 30 Nov. 2001 |
| Prof. C.A.B. Smith | 1917 | 1963 | 10 Jan. 2002 |
| Prof. Jacques A. Zighera | 1929 | 1994 | Aug. 2001 |
Carl-Gustav Esseen (September 18, 1918 - November 10, 2001)
Professor Carl-Gustav Esseen grew up in Linköping, where he finished high school with top grades. In the fall of 1936 he came to Uppsala to study mathematics, astronomy, physics and chemistry. His main interest since his early days, however, was mathematics, in particular the calculus of probabilities. He obtained his Ph.D. degree in the fall of 1944 with an extensive and deep thesis entitled "Fourier analysis of distribution functions. A mathematical study of the Laplace-Gaussian law". His thesis was published in Acta Mathematica, one of the most prestigious mathematics journals.
The normal distribution, or the Gauss distribution, is a classical concept in probability theory. The distribution originates from the fact that the cumulative effect of independent trials can be approximated by the normal distribution - this is the Central Limit Theorem. Inspired by the works of Harald Cramér and the legendary Uppsala mathematician Arne Beurling, Carl-Gustav Esseen investigated the accuracy of the approximation. By ingenious Fourier analytic computations he obtained the sharpest possible estimates of the approximation error; he solved the simplest case, when all the summands are independent and have the same distribution, as early as the fall of 1940 when he was only 22 years old. The key step in his proof is now known as the Esseen Lemma, and it is an inequality in which the difference between two distribution functions is estimated via an integral based on the difference between the corresponding characteristic functions. The method and technique established by Esseen, and epitomised by this renowned lemma, have been exploited by researchers throughout the world ever since.
After a few years as Assistant Professor in Uppsala, Carl-Gustav Esseen was appointed Full Professor of Applied Mathematics at the Royal Institute of Technology in Stockholm. In those days a full professor, also known as `The Chair', was responsible for all kinds of mathematics. As a consequence of the rapid development of Mathematical Foundations as well as the technical applications, however, the professorship was split into two: One in Information Technology, and in particular numerical analysis; and one in Mathematical Statistics. In 1962 Esseen moved to a Chair in Mathematical Statistics. He was subsequently elected a member of the Royal Swedish Academy of Engineering Sciences in 1963.
In 1967, after 18 years at the Royal Institute of Technology, Carl-Gustav Esseen became the first holder of the Chair in Mathematical Statistics at Uppsala University, a position he held until his retirement at the end of 1984.
During his many years as a professor Carl-Gustav Esseen guided several PhD students to their degrees. Throughout the years he was also known for his meticulously prepared and perfectly delivered lectures on a wide variety of topics. His lecture notes in several areas have been influential for innovations in Swedish industry, for example in Control Theory and in Telecommunications.
The wide scope of knowledge Carl-Gustav Esseen demonstrated was impressive. On several occasions the problems he presented from books and papers had been insufficiently investigated, on other occasions proofs were missing or incorrect. In his modest manner he would comment on this with phrases such as "I had to fix that'" or "there was no proof, but it turned out that it was possible to prove the result as follows", which typically was not at all an easy task, but rather a contribution to the field. Had he wanted, many learned books and papers could have been written by him. Instead his passion was simply to discover for its own sake, and to teach. In addition, an internal aesthetic that required perfection to his own standards, he kept his list of publications short, but always of the highest quality.
He remained faithful to his department after his retirement, until health and eyesight began to fail. During this period he returned to one of his irst intellectual loves, that of number theory, where he investigated problems concerning the factorisation of large numbers, a topic that has become important in cryptology.
Carl-Gustav Esseen was by no means a man of gestures and ceremonies. There is a Swedish saying to the effect that great achievements are accomplished quietly. A great master, one of our important scientists has left us. We shall miss him.
Allan Gut
Professor of Mathematical Statistics at Uppsala University
Lars Holst
Professor of Mathematical Statistics at The Royal Institute of Technology
Shanti Swarup Gupta (January 25, 1925 - January 11, 2002)
Dr. Gupta was born in Saunasi, Mainpuri, India, on January 25, 1925. After completing a master's degree in mathematics at the University of Delhi and several years of teaching there, he came to the United States as a Fulbright Scholar in 1953. He received a doctorate in statistics from the University of North Carolina at Chapel Hill in 1956 and subsequently worked in statistics research at Bell Labs in Allentown, PA.
Dr. Gupta held teaching positions at the University of Alberta and the Courant Institute, New York University, as well as visiting appointments at Stanford University, the University of California-Berkeley, and the Technical University of Darmstadt, Germany.
In 1968 he founded the Purdue University Department of Statistics and served as department head until 1995. During this period and up to his illness, he was engaged in research and teaching. He was a Professor of Statistics and Mathematics at Purdue University from 1962 until his death.
His honours include Fellow of the American Association for the Advancement of Science, of the American Statistical Association, of the Institute of Mathematical Statistics, as well as Erskine Fellow of the University of Canterbury in New Zealand, Honorary Fellow of the Indian Statistical Association, Special Chair in Academia Sinica, Taipei.
Dr. Gupta served as President of the Institute of Mathematical Statistics and Editor-in-Chief of the Journal of Statistical Planning and Inference. He was a prominent scholar in the area of statistical ranking and selection, and authored more than 200 published articles and books and was the major advisor to 30 doctoral students.
Dr. Gupta is survived by his wife Marianne Heinicke Gupta, and their daughter Maya Gupta.
Courtesy of Purdue Department of Statistics
http://www.stat.purdue.edu/
Henry Oliver Lancaster (1 February 1913 - 2 December 2001)
Oliver Lancaster's life as student and scholar was spent at Sydney University
where he initially enrolled in 1930. During this time he achieved scholarly
distinction in at least four fields: mathematical statistics, medical and public
health statistics, the history of medicine and of statistical theory, and
statistical bibliography. His early career was with the School of Public Health
and Tropical Medicine (S.P.H.T.M.), from which he was appointed to the Foundation
Chair of Mathematical Statistics in 1959, a position he held until his retirement
in 1978. His achievements on the medical side included definitive verification in
1951 of the connection between rubella in early pregnancy of the mother and
congenital deafness of offspring.
He was much honoured in his lifetime. Elected Fellow of the Australian Academy
of Science (FAA) in 1961, he was awarded its Thomas Ranken Lyle Medal for
Mathematics and Physics in the same year. He was one of the early recipients
of the Pitman Medal of the Statistical Society of Australia for distinction
in research in 1980. He was made Officer of the Order of Australia in 1992.
Henry Oliver Lancaster (H.O.L.) was born the second son of Dr. Llewellyn Bentley Lancaster, a medical graduate of Sydney University, and his wife Edith Hulda Smith during a visit to Sydney, and spent his early years in Kempsey, NSW, where his father practised. From 1940 he worked as Medical Officer (Pathologist) in the Australian Imperial Forces, first in the Middle East, then from early 1942 in the 117 Australian General Hospital in Townsville, this laying the groundwork for his first (joint) papers which appeared in the Medical Journal of Australia in 1944. Here he encountered the problem of analysing a 2x2x2 table of frequencies. Service in New Guinea followed, and 1944 saw his secondment to the Australian New Guinea Administrative Unit, and with it an awakening of interest in demography, and a return to the serious study of mathematics, for which he had shown extraordinary aptitude in his early years.
Given a temporary appointment in 1946 at the S.P.T.H.M. as Lecturer in Medical Statistics, he spent much of his time completing his mathematics education and reading the English and American statistician-epidemiologists. Bradford Hill had succeeded Major Greenwood as Professor of Statistics at the London School of Hygiene when H.O.L. arrived there in 1948 as Rockefeller Fellow in Medicine. He shared an office with Peter Armitage, who was to become an outstanding biostatistician, a President of the Royal Statistical Society, and a friend for life. The major influence on H.O.L. at the School of Hygiene, however, was J.O. Irwin (1898-1982) who encouraged H.O.L. with his first paper in mathematical statistics, published in 1949, and in continuing work on partitions of chi-squared statistics. This advice H.O.L. followed, with a Ph.D. from Sydney University in 1953 and a series of publications highlighted by his book The Chi-squared Distribution (1969).
The driving forces behind the inaugural meeting of the Statistical Society of New South Wales, held at the University of Sydney on 25 September 1947, were H.O.L., Helen Newton-Turner (who became first President), and R.S.G. Rutherford (later Professor of Economic Statistics). In these early years the Society included a number of mathematical statisticians who achieved prominence overseas, such as H.A. David and D.B. Duncan. On his return from London, H.O.L. was principally responsible for the publication of the Bulletin of the Statistical Society of New South Wales which was replaced in 1959 by the Australian Journal of Statistics with H.O.L. as founding editor. He served continuously till 1971. The name for the new journal anticipated there soon being a Statistical Society of Australia, whose nucleus was the amalgamation of the N.S.W. and Canberra Statistical Societies in 1962.
The first 4th year Honours class in Mathematical Statistics was in 1960, with M. Aitkin, R.B. Armson, B. Bennett, C.C. Heyde, and M.A. Hamdan. Several went on to distinguished academic careers in mathematical statistics, as did the Ph.D. students H.O.L. supervised, the first of which were G.K. Eagleson and M.A. Hamdan.
H.O.L's work on stochastic dependence went in particular from chi-square to the study of bivariate distributions and their spectral expansions in terms of correlations and orthogonal polynomial systems. His very successful treatment of the Meixner class (J.R.S.S.(B) 37,434-443) was among his favourite 7 papers of his "mathematical" statistics stream, among which his first choice was on his characterisation of the normal distribution (J. Austral. Math. Soc., 1 (1960) 368-383). Another of his favourites (J. Amer. Statist. Assoc., 56 (1961) 223-234) lays the foundation for what has come to be known as Lancaster's mid-P adjustment.
The 1967 meeting of the ISI was in Sydney, and brought many distinguished overseas visitors to Australia for the first time. H.O.L.'s Department of Mathematical Statistics played a key role in hosting the visitors, senior students being engaged to assist. I remember a function in H.O.L's spacious office where, as a young doctoral student from Canberra and friend of Geoff Eagleson's, I was present, but blundered when D. Dugué wanted to make a note, by handing him a card from H.O.L.'s precious bibliographic index cards. Dugué noticed the writing on the other side, and Geoff quickly restored it to its pile, without (fortunately for our future good relationship) H.O.L. noticing. H.O.L. was a long-term member of the ISI, (elected 1961). He played an active part in the London Meeting of the ISI in 1969, during a year (1969-1970) spent in Great Britain.
H.O.L.'s scholarly activity of bringing together a large bibliography of statistics was motivated by the belief that authors don't give enough credit to earlier work laying foundation to their own. This was associated, on the one hand, with his interest in the history of statistics which motivated historiographic activity by C.C. Heyde and the undersigned (occasionally one heard of an Australian historical "School"); and on the other (at the invitation of the ISI) to the Bibliography of Statistical Bibliographies (1968), with its 21 Addenda over the years to 1989. This celebrated activity was recognised by the Royal Statistical Society before which he read an invited paper entitled "Problems in the Bibliography of Statistics" on Wed April 8, 1970 (J.R.S.S. (A) 133, 409-441). The discussion of this paper and one of J. Gani published immediately following stretches over pp. 450-462! I remember in particular a vivid statement by H.O.L. on this occasion, that - due to factual inaccuracies perpetrated by biographers - every statistician should write his own obituary. I had cause, as his successor, to remind him often of this statement during his retirement. For a time he resisted, saying that it would be the job of the person given the task to do it properly. An autobiography appeared in The Making of Statisticians (Ed. J. Gani, 1982, pp. 235-252), and eventually, under the additional influence of his brother, this was expanded into Some Recollections of Henry Oliver Lancaster (ed. R.L. Lancaster, 1996, primarily printed for the Sydney University Archives). He concludes it with: "A continued interest in some general problem of the world at large, related to an individual's technical expertise, can be its own reward." This is a humble epitaph, for a great scholar.
Eugene Seneta
Harold Ruben (April 3, 1923 - November 30, 2001)
Harold Ruben was born on April 3, 1923, in Rawa, Poland, and died in Harlow, Essex, England, on November 30, 2001. The family moved to England when Harold was quite young. For both his undergraduate and graduate studies, Harold went to Imperial College London. He completed a Bachelor of Science degree with honours in mathematics in 1943 and became an Associate of the Royal College of Science; his Doctor of Philosophy degree in mathematical statistics was awarded in 1950.
Harold Ruben's academic career began in 1950 with a three-year Senior Research
Fellowship from the Agricultural Research Council. He went first to Aberdeen,
Scotland, and then to the Department of Genetics at Cambridge, where he worked
with Lord Rothschild and M.M. Swann. [Nathaniel Mayer Victor Rothschild,
GBE, GM, FRS (1910-1990), was Chairman of the Agricultural Research Council,
1948-1958. In 1946 Michael Meredith Swann, FRS (1920-1990), went to Cambridge,
where "he became a considerable authority on cell biology" [The Times,
24 September 1990]; Swann was chairman of the BBC 1973-1978.]
In 1953 Harold became Senior Lecturer at the University of Manchester. Four years later he moved to North America as a Visiting Professor in the Department of Mathematical Statistics at Columbia University in New York City.
In 1960 Harold returned to England as Head of the Department of Statistics at the University of Sheffield. Three years later he was again back in North America, this time as Professor in the Department of Statistics at the University of Minnesota. Lecture notes prepared from courses Harold taught at Minnesota were used for many years as the "text" for the core graduate course on "Mathematical Statistics" at McGill. At Minnesota, Harold met Charles Kraft and Constance van Eeden, who both later joined the Département de mathématiques at the Université de Montréal in 1965.
In 1966 Harold returned to England again, this time becoming Head of the Department of Statistics at Birkbeck College London; in 1969, encouraged by Charles Kraft and Constance van Eeden, Harold moved to Montreal as Professor in the Department of Mathematics at McGill University, where he stayed until his retirement at the age of 65 in 1988, just before mandatory retirement ended at McGill. Harold moved back to England, where he stayed, living most recently in Epping, Essex (near London). Harold was married and his wife Jean survives him; there were no children.
Harold spent four summers (1959, 1960, 1962, 1965) as a Visiting Professor in the Department of Statistics at Stanford University and was also a Visiting Consultant at Boeing Research Laboratories in Seattle, Washington, from 1959 through 1969. At Boeing, he met George Marsaglia, who moved to Montreal to become the first Director of the School of Computer Science at McGill University in 1970. Harold Ruben became a Fellow of the Royal Statistical Society (RSS) in 1948 and served on the RSS Council; he was elected an Ordinary Member of the ISI and a Fellow of the Institute of Mathematical Statistics (IMS). He was also a member of the London Mathematical Society.
Harold Ruben spent almost twenty years (1969-1988) at McGill, and for all this time he was the senior member of the statistics group in McGill's Department of Mathematics and Statistics (named Department of Mathematics until 1983). Harold supervised five M.Sc. theses, all at McGill, and three Ph.D. dissertations--all but one at McGill: Gordon Simons completed his Ph.D. under Harold's guidance at Minnesota in 1966 and is presently a Professor in the Department of Statistics at the University of North Carolina at Chapel Hill. Philip McDunnough is a Professor in the Department of Statistics at the University of Toronto and Louis-Paul Rivest is Professeur in the Département de mathématiques et de statistique at Université Laval in Québec City. Harold's five M.Sc. students Claude Bélisle, Michael Edwardes, Jack Gambino, Bill Reed, and Jack Siemiatycki all went on to complete Ph.D. degrees in statistics.
The courses that Harold Ruben taught were well-motivated, meticulously organised, conducted with spirited zeal, and his lectures were rich in historical perspective. His students recall his tireless enthusiasm, even though by the later years he had taught the same material dozens of times. The examples he gave were always interesting, with those based on RAF bombing raids during Word War II particularly memorable. We are saddened to think that we will no longer have an opportunity to attend a Ruben lecture.
Harold's research publications are all in mathematical statistics and related topics. His first contribution was a statistical appendix to a paper in experimental biology by Rothschild and Swann published in 1952; in this appendix Harold "devised a rigorous way of analysing the experimental data". Another early contribution was on the problem of measurement in the psychological sciences, jointly authored with Wolfe Mays. Harold's major contributions to mathematical statistics, however, are in multivariate statistical analysis with strong connections to geometric probability and statistical physics. An important early paper in geometric probability was published in the prestigious Acta Mathematica. Harold's work in geometric probability continued in the 1970s with a series of papers. Almost all of his publications were in leading statistical journals such as The Annals of Mathematical Statistics, Biometrika, Journal of the Royal Statistical Society, Series B: Methodological, and Sankhya: The Indian Journal of Statistics, Series A. Of particular note is the series of four papers on the "Probability content of regions under spherical normal distributions'' published in the Annals in 1960 through 1962. Another series of four papers was on "Studentisation of two-stage sample means from normal populations'' published in Sankhya at almost exactly the same time. And yet another series of four papers were on Mills's ratio (the reciprocal of the hazard rate or failure rate) used in approximating the tail of a distribution and named after J. F. Mills [Biometrika, 18 (1926), 395-400].
Harold Ruben's papers are all very well-written as his way with words was very expressive, and his scholarship, admirable; this is particularly evident in his book reviews. In one review Harold observed that ``... one cannot but be impressed by the author's achievement--a veritable tour de force--in encompassing, with erudition and in a spirit of scientific catholocity but yet succinctly and in compact form, a mass of general results with a wealth of diverse applications in so narrow a space''. And Harold did not withhold criticism: in another review Harold wrote: "I fear that after struggling with this book the earnest but uninitiated reader will, like the Persian poet-philosopher, come out by the same door as in he went. Not a few statistics books have appeared recently which are insufficiently motivated, poorly organised, substantively thin and bereft of all cultural depth or historical perspective.''
Harold Ruben was a staunch defender of the view that statistics and mathematics were two completely different disciplines, driven by different motivations. Some of his mathematician colleagues found this puzzling especially as Harold had more than a passing interest in pure mathematics, having published a long paper in Acta Mathematica; when he wore his statistician's hat, however, he regarded mathematics merely as a tool. For years he advocated that statisticians should reside in a different department from the mathematicians, which is now the norm.
Harold Ruben will be remembered not only as a statistician and mathematician whose contributions are widely recognised, but also as an intellectual whose interests were broad. He held strong views on many subjects, which he would argue, often with vehemence, in the Department Lounge at McGill. Harold was impatient with the practicalities of administration. He represented the learned class of well-read academics who care about science, about literature, about politics, and about world affairs. He was, in short, a true scientist.
His students and friends also cherish warm memories of Harold the man: his unconscious casual dress, the chalk dust rendering his spectacles almost opaque during class, the soft chuckle of his laughter, that he always drank afternoon tea with lemon and from a glass, and how he enjoyed smoking cigars (sometimes big, fat ones). He had a very sharp mind as well as a sense of humour. Honest, kind, open, and outspoken, he was also a very private person.
A complete list of Harold Ruben's research publications, as well as the titles of the theses and dissertations of Harold's Ph.D. and M.Sc. students, is posted on my home page http://www.math.mcgill.ca/~styan/
I am very grateful to Mrs. Jean Ruben for her help in completing this obituary, as well as the following: T. W. Anderson, Mira Baksalary, George A. Barnard, Daniel Berze, Arup Bose, William G. Brown, Theophilos Cacoullos, Donald A. Dawson, Michael D. deB. Edwardes, Jack Gambino, Christian Genest, Lawrence Joseph, Wolfe Mays, William J. Reed, Louis-Paul Rivest, Georg Schmidt, Hilde Schroeder, Vanamamalai Seshadri, Milton Sobel, Evelyn Matheson Styan, Elaine Tremblay, Constance van Eeden, Christina M. Wolfson, and David B. Wolfson.
George P. H. Styan, McGill University, Montreal
Back to Home
Back to Newsletter
